Continuous Simulation
Geological properties are variable at all scales and there is variability/heterogeneity at smaller scales than we can resolve with the available data. This variability can have an effect on engineering design and resource calculations. Simulation creates realizations of what the variability might look like. Simulation is also used for uncertainty quantification, but greater care must be taken for this goal.
Almost always, local uncertainty is calculated under a statistical model such as the multivariate Gaussian model or the direct estimation of conditional probabilities with, say, the indicator model. In some cases, with object-based modeling and with multiple point statistics for example, local uncertainty can only be understood by creating multiple realizations. The use of simulation is twofold: create realizations that show reasonable patterns of variability and provide a model of uncertainty at a larger scale. These lessons focus on the simulation of continuous variables, but there is some overlap with the simulation of categorical variables.
Lessons
- Trend Modeling and Modeling with a Trend (see source code on GitHub)
- Cell Declustering Parameter Selection
- Aggregating Variables into a Super Secondary Variable
- Transforming Data to a Gaussian Distribution
- The Decision of Stationarity
- Choosing the Discretization Level for Block Property Estimation
- Multivariate Gaussian Distribution
- Localization of Probabilistic Resource Models
- Checking Continuous Variable Realizations - Mining